Summary The crew of a Transport Canada Beech King Air C90A, registration C-FGXU, serial numberLJ1140, departed Winnipeg, Manitoba, at 0940 central standard time1 on a routine training flight to Prince Albert, Saskatchewan. The captain was the pilot flying and the co-pilot was the pilot non-flying. At approximately 1026, at flight level 220, the crew heard a loud bang accompanied by severe airframe vibration and a substantial pitch up in aircraft attitude. The captain disconnected the auto-pilot, reduced engine power, selected full nose-down trim, and applied forward pressure on the control column to regain control. As the airspeed reduced to below approximately 150knots, the vibrations stopped. With limited elevator control remaining, the captain reduced engine power and established a descent while maintaining hard nose-down pressure on the control column to keep a constant level aircraft attitude. The co-pilot declared an emergency and requested a diversion to Dauphin, the nearest suitable airport for landing. At 15000feet, the captain elected to extend the landing gear to effect a more rapid descent and as a precaution to give more time to stabilize the aircraft in the event that lowering of the landing gear would affect airflow past the tail. Prior to landing, at approximately 200feet above ground level, the crew detected another brief vibration followed by a sudden pitch down, requiring aggressive trim and elevator control inputs to control. The crew completed a flapless landing without further incident. After exiting the aircraft, the crew observed that the left elevator trim-tab pushrod had failed. Ce rapport est galement disponible en franais. Other Factual Information Transport Canada (TC) documentation indicated that the crew was certified and qualified to perform the flight crew duties they were assigned. Prior to departure, the pilot-in-command operated all of the flight-control trims through their full range of movement, and both crew members checked the elevator electric pitch trim functions in accordance with standard operating procedures. The elevator trim checks were normal; no binding or excess force was required for operation. The elevator trim system incorporates a pushrod attached to the trim tab with a set of stainless steel bushings, an inner bushing rotating inside an outer bushing, at the attachment point. The last major inspection of the elevator trim system was accomplished during a phase2 inspection of the aircraft on 21October2002, at an airframe time of 7901hours. The elevator trim-tab free play was checked and found to be within the required manufacturer's specifications. The elevator trim-tab pushrod attachment hardware was removed to facilitate the installation of new magna-fluxed hardware (MS17825-4 castellated fibre locknuts and close tolerance AN174-7bolts). After removal of the hardware, it was noted that the steel inner bushing (p/n90-524024-1) on the left trim-tab arm was seized within the steel outer bushing (p/n90-610010-5). The inner bushing was removed and found to be corroded and rusted. Both mating faces of the inner and outer bushings were cleaned, and the bushings were lubricated and re-installed with the new attachment hardware. The elevator trim was run through its full operating range and was found to operate smoothly, with no signs of binding. Photo1. Failed left elevator trim-tab pushrod The occurrence happened approximately 150flight hours after the inspection, at an airframe time of 8050hours. An examination of the aircraft after the occurrence revealed that the left elevator trim-tab pushrod had failed at the threaded rod, flush with the surface of the jam nut, adjacent to the adjustable fork-end (seePhoto1). The fork-end was seized at a right angle to the trim-tab horn. The pushrod cutout in the elevator lower surface showed rub marks from the jamming of the fork-end in several locations in the cutout after the failure occurred. The torque on the attachment hardware was checked and found to be 90inch-pounds. The manufacturer's recommended torque is 25to 30inch-pounds, with a provision to go as high as 50inch-pounds to aid in aligning the slot in the castellated nut. It was reported that a torque wrench had been used during the installation of the castellated nut. The trim tab and attachment hardware was forwarded to the Transportation Safety Board of Canada (TSB) Engineering Branch for examination. Scanning electron microscope examination of the fracture face on the threaded rod showed fatigue originating in a thread root. The fatigue cracking had progressed to overstress rupture with no noticeable final overstress region, suggesting that the failure occurred under normal service loads when the crack reached critical size. Metallurgical testing of the pushrod showed that it met material specifications. Examination of the inner faces of the trim pushrod clevis showed a worn recessed area and rub marks around the bore for the clevis/horn attachment bolt. Examination of the horn assembly showed that mechanical rubbing had removed the paint from the horn around the periphery of the outer bushing, consistent with contact between the sides of the horn and the trim pushrod clevis. Photo2. Inner and outer bushing arrangement The inner bushing was removed and a significant amount of non-uniform, circumferential galling2 was observed extending around the entire circumference of the outer surface of the inner bushing. A dimensional check against the manufacturer's specifications showed that the inner bore of the outer bushing was undersized and non-cylindrical, while the diameter of the inner bushing was oversized, resulting in an interference-fit3 between the two bushings and the non-uniform galling. The manufacturer's designed clearance range is between 0.0001and 0.0016inches; the measured clearance was between 0.0000and -0.0013inches. TC inspected two new inner bushings and one new outer bushing from stock that had been supplied by the aircraft manufacturer. A dimensional check of the bushings showed that they were non-cylindrical in nature (out-of-round) and outside of the manufacturer's dimensional specifications. The two inner bushings were between 0.0003and 0.0004inches oversized, the outer bushing was 0.0009undersized and all three bushings were between 0.0006and 0.0008inches out-of-round. There are no out-of-round limits. The oval nature of the bushings, if aligned, would have resulted in an interference fit and would not have allowed the inner bushing to rotate. The TSB Engineering Branch recently examined a failed trim-tab pushrod from a commercial operator's Beech King Air90 aircraft after a similar occurrence on 05June2002. The failure mode was similar to the failure mode in this occurrence.